← Back to community benchmarks

Qwen3.5-122B-A10B-oQ4

M5 Max (40c) · 128 GB · 4bit · 2026-03-28
Performance
16k
tokens
817.2
PP tok/s
13.9
TG tok/s
20048
TTFT (ms)
75.8
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.2.24.dev2
macOS macOS 26.4
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
4k 1,078 15.1 72.1 GB view
8k 966.4 14.5 73.3 GB view
16k 817.2 13.9 75.8 GB current
32k 693.9 12.4 80.6 GB view